Optimising reporting for AI / Section 7:
What lies ahead?
With AI here to stay, optimising reporting content for AI will become increasingly important and traditional processes for creating and designing reports will be replaced.
In just a few years, AI and digitisation have transformed audiences’ access to information and this is already shaping their decisions.
Contrasting paces of change
In 1844, the Joint Stock Companies Act was introduced by William Gladstone and corporate reporting in the UK was formalised. Much has changed since then, but the Act’s details with its references to balance sheets, reporting deadlines, auditors and presentation of accounts at General Meetings indicate a strong common thread has been retained over the subsequent 180 years. Corporate reporting requirements have changed in many ways, but these changes have usually been incremental, and relatively slow.
The near future for corporate reporting can often be predicted with confidence, in part because changes are usually signposted by regulators well in advance. At the time of writing, there is a consultation relating to the introduction of sustainability reporting standards to corporate reporting for UK listed companies. We anticipate that this proposal will proceed and if so, a key element of the next four or five years of change to corporate reporting will have been mapped out for all who are interested.
AI, in contrast, is driving change at a much faster pace and this makes predicting the future a little harder. But because digitisation of corporate reporting processes has been relatively slow to arrive, predicting its future has one particular advantage. Simply, that many other disciplines and markets, digital transformation has already happened. Parallels can be drawn.
There is one common thread to all changes that digitisation has brought – they have all been led by changes seen in end users, customers or audience behaviour.
Whether in retail, entertainment, education, music, or travel, it has been the choices made by consumers that accelerate change. Whenever a better, more efficient, cheaper or more convenient digital option became available, these digital services replaced the process that had previously delivered goods or services. This provides a good indication for predicting what is most likely to happen in reporting – much of this can already be seen in changed user behaviour.
Where digitisation offers any advantage for investors and stakeholders, this is what they will expect, choose and perhaps insist upon.
AI and digitisation are inevitably going to reshape reporting. The key question is not if, but when, and how?
Two framing questions which help when predicting the future
If you are in any doubt about the future of reporting, try asking yourself the following questions. We suspect you will find the answers straightforward:
Do you think reporting will forever remain as it is today, with all versions created PDF-first using print and document-based software?
Or, do you think reporting will eventually shift to digital-first processes and outputs that AI tools can reliably read?
At Reportl, we made a decision some years ago that AI and digitisation will inevitably reshape reporting. The key question is not if, but when, and how?
For the answers to this, we look again at reporting today, the changes that AI brings and regulators’ responses to demand.
The print-first PDF format is reliable and well-liked by humans. But it has a fundamental and unsolvable flaw. PDF is not a reliable source of content and data for AI to read accurately.
It also has some well-known practical limitations. It can be almost impossible to read on a mobile, and most PDFs are inaccessible to digital tools like screen readers and search engines.
Today’s user data suggests that PDF will still remain a vital output format for humans. But it offers no new potential as the future ‘source’ format for reporting.
So something needs to change. And, fortunately this change is already underway.
Mandatory digitisation began a few years ago with the ESEF report format. This web format (HTML) is now required for reporting in most major markets.
Don’t be deterred by the unusable ESEF reports you know today. Most of these are limited by the PDFs they are converted from. The translation of key disclosures into code that machines can read must be part of the future and it can be harmoniously blended with the design and unfurling narrative that people want to see in a corporate report.
The future of ESEF is the future of reporting.
This may be appear to be a bold claim. But when you see the full power of this format we think you will agree.
When created digital-first (not converted from PDF) the mandatory ESEF web format meets every conceivable need. It can be optimised for AI, Google-friendly, accessible and mobile responsive. It can also be tagged for instant and accurate machine analysis, and can be read online by anyone with a browser – no special software required.
Another remarkable capability of ESEF HTML code which many are unaware of today… it can also be used to publish a PDF and print report that is as beautiful and engaging as any created in print-first software.
So which process do you think is most likely to lead reporting into the future? Traditional print-first? Or digital-first? In this context, whilst nothing can be guaranteed, we offer three connected predictions about the future of reporting:
When technology quickly changes people’s behaviour, or its future lifespan appears long lasting, organisations adapt very quickly.
AI, it has already changed people’s behaviour.
Prediction #1:
AI is here to stay
Key takeaways:
AI is becoming a ‘foundational’ tech, like electricity or the internet.
AI skills and expertise will become an embedded element of successful teams.
In the early days of the World Wide Web there were many who thought it was a fad, a short-term gimmick. Even in the year 2000, some newspaper reports suggested that the internet’s best days were over. This may have been wishful thinking, fearful of its impact on their sales.
But maybe not, because without the benefit of hindsight, for any exciting but uncertain technological advance, a sensible question to ask is “Will this become obsolete? If so, when?”
Much technology gets superseded eventually. But when technology quickly changes people’s behaviour, or its future lifespan appears long lasting, organisations adapt very quickly – usually to meet new demands or exploit new opportunities.
In the case of AI, it has already changed people’s behaviour. And it is clear it will become a long-term ‘foundational’ technology, rather like electricity networks or the internet itself (which wasn’t a fad after all).
The last few years have seen a frenzy of interest in AI, with echoes of the dot com bubble of the late nineties. Many AI applications may arrive with great fanfare, flourish before being surpassed. Fortunes will be made and lost.
But while specific applications using AI may come and go, the core technology and capabilities of LLMs will continue to evolve and change many aspects of life and work, in some cases at a fundamental level. Even at this early stage, we can all be confident that AI technology is here to stay.
In the same way that website, social media, and e-commerce teams have become a part of most organisations, it seems obvious that AI will also bring new strategies, skills and expertise to teams – they will be needed to deliver best-in-class AI-related services and processes for years to come.
And like the web, AI will challenge many print and PDF conventions around reporting. Not least the obsession with page counts, which become less relevant in a digital and AI-led world where rich, authoritative detail matters just as much as clarity of message.
But beyond new skills and new applications that use AI, we think its effects will change some industries at a more fundamental level. One of these is corporate reporting.
And this is the basis of our second prediction.
AI ignores PDFs and looks for other sources online.
Prediction #2:
All reporting content
will be optimised for AI
Key takeaways:
AI users will demand optimised online reports that AI can read accurately, not PDFs.
Companies will eventually want more tagging so their content is read properly and served up by AI.
Everyone has experienced AI’s biggest flaw – hallucinations. But most AI errors are caused by the lack of high-quality digital sources, not AI itself. Garbage in, garbage out.
AI accuracy suffers when it only has PDF sources. AI frequently ignores PDFs and looks for other sources online. As investors’ reliance on AI tools increases, more precise, trusted and findable reporting from companies becomes essential.
Therefore, we think that all reporting is going to need to be optimised for AI – so that any AI tool can find, read and analyse all reporting content and data accurately.
Regulators introduced the perfect solution – the online ESEF report.
Understanding of reporting is changing
AI’s demand for digital data is shifting people’s understanding of what reporting is, from ‘document’ based reports (not AI-friendly), to digital reports that are mini online ‘databases’.
Recognising the limitations of PDF as a data source, regulators introduced a highly-effective solution, the digital ESEF report. This can provide accessible data for AI tools to crawl and enable faster, and more accurate analysis for every stakeholder.
Try asking a simple question in your favourite AI tool: “How do I optimise my annual report format for AI?” The AI tool will probably tell you the importance of using HTML, not PDF, and it may even explain why it is helpful to also tag this content with XBRL and/or JSON-LD.
This optimisation is best delivered in the online ESEF format: HTML with tagging.
To be visible to tools like ChatGPT, the report must be created from HTML (not converted from a PDF). Tagging the report gives AI the precise meaning it needs to deliver consistent, reliable answers and analysis. This is why regulators recommend improving your ESEF report.
Now that AI is becoming the first access point for content and analysis, it seems inevitable that companies will want their reporting to be tagged for AI to find and read.
Will demand for tagging on reports increase?
To help answer this, it’s useful to look at what happened on the web. As search engines like Google became the main access point, ‘meta tagging’ became a top priority on websites. This spawned a global industry – SEO.
Now that AI is becoming the primary access point for content and analysis, it seems inevitable that companies will want their reporting to be tagged for AI to find and read. Increased demand may also come from regulators and ESMA has proposed detailed notes tagging because this would significantly improve AI accuracy.
So, we predict that AI will increase understanding of tagging, and accelerate a shift that has already begun. All reporting will eventually be designed and built from native HTML, tagged, and published in full online, optimising the information for AI tools (GEO).
We expect this will reach a tipping point when the question “Digital-first reporting: Why should we do that?” becomes “Digital-first reporting: Why aren’t we doing that?”
How it will be achieved is our third prediction.
A gold standard is already emerging with next generation digital platforms.
Alongside the PDF version is an AI-readable, fully-designed online report with identical content, which is tagged for maximum AI accuracy.
Prediction #3:
Reporting design processes
will need to digitise
Key takeaways:
Companies are realising PDF-first data compromises AI’s value.
Design teams will move to digital-first (native HTML) design software, replacing their print-first tools.
Most discussion about AI’s future in reporting focuses on how it may or may not be used to help collate, generate, refine and verify content. Good governance over AI usage is going to be vital. However, we think AI will also bring another fundamental change to reporting.
Continuing to create reports using a PDF-first process increases the risks of AI making errors. Companies are going to demand solutions that reduce these risks.
This can already be seen in best practice digital-first ESEF reporting. A gold standard is already emerging with next generation digital platforms like Reportl, which deliver reports from HTML at source, not from PDFs. As well as bringing more accurate data for AI, this is also bringing other enhancements like search optimisation, accessibility, mobile responsiveness and increases in visitors.
Such digital-first reporting platforms solve AI readability very simply – by reversing how a report is produced. No longer created using PDF-first ‘print’ code as the source, in Reportl all formats are built from one HTML source, including the PDF version and the online report.
By making this shift, PDF becomes just one of the outputs, no longer the primary data source. And alongside the PDF version is an AI-readable, fully-designed online report with identical content, which is tagged for maximum AI accuracy. Processes will become smoother, with late nights perhaps becoming rarer, as late-stage sequential processes and controls are streamlined by the change.
At the start of this white paper we promised not to ‘sell’ Reportl. Whilst it is built to meet the fast-changing reporting demands which this white paper highlights, there are several other excellent platforms that also offer the benefits of digital-first reporting. Each of these serve slightly different needs in the market, and we are happy to recommend alternative solutions where they may better meet your needs.
If our predictions, or any other aspect of this white paper interests you, the team at Reportl would be very happy to hear from you at [email protected]